Transductive Multi-View Zero-Shot Learning
Open accessThe DOI matches its Crossref record; 72 of the 74 references it cites were found in a registry or an index.
Abstract
Most existing zero-shot learning approaches exploit transfer learning via an intermediate semantic representation shared between an annotated auxiliary dataset and a target dataset with different classes and no annotation. A projection from a low-level feature space to the semantic representation space is learned from the auxiliary dataset and applied without adaptation to the target dataset. In this paper we identify two inherent limitations with these approaches. First, due to having disjoint and potentially unrelated classes, the projection functions learned from the auxiliary dataset/domain are biased when applied directly to the target dataset/domain. We call this problem the projection domain shift problem and propose a novel framework, transductive multi-view embedding, to solve it. The second limitation is the prototype sparsity problem which refers to the fact that for each target class, only a single prototype is available for zero-shot learning given a semantic representation. To overcome this problem, a novel heterogeneous multi-view hypergraph label propagation method is formulated for zero-shot learning in the transductive embedding space. It effectively exploits the complementary information offered by different semantic representations and takes advantage of the manifold structures of multiple representation spaces in a coherent manner. We demonstrate through extensive experiments that the proposed approach (1) rectifies the projection shift between the auxiliary and target domains, (2) exploits the complementarity of multiple semantic representations, (3) significantly outperforms existing methods for both zero-shot and N-shot recognition on three image and video benchmark datasets, and (4) enables novel cross-view annotation tasks.
Overview
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cites
ImageNet classification with deep convolutional neural networks
Krizhevsky, A., Sutskever, I., & Hinton, G. E. (2017) · Communications of the ACM · doi:10.1145/3065386 · Cited by 112,155 · cited 1×
“What is interesting to see here is that Overfeat > Decaf since both are based on the same Convolutional Neural Network (CNN) model of [32].”
BackgroundCollective dynamics of ‘small-world’ networks
Watts, D. J., & Strogatz, S. H. (1998) · Nature · doi:10.1038/30918 · Cited by 43,060
LIBSVM: A library for support vector machines
Chang, C.-C., & Lin, C.-J. (2011) · ACM Transactions on Intelligent Systems and Technology · doi:10.1145/1961189.1961199 · Cited by 40,918 · cited 1×
“However, this does not effectively exploit the multi-view complementarity, and suffers from labelled data (prototype) sparsity.”
Background(2008) · Journal of Machine Learning Research · Cited by 33,051
Pan, S. J., & Yang, Q. (2010) · IEEE Transactions on Knowledge and Data Engineering · doi:10.1109/tkde.2009.191 · Cited by 23,387
Learning Deep Architectures for AI
Bengio, Y. (2009) · Foundations and Trends® in Machine Learning · doi:10.1561/2200000006 · Cited by 9,239 · cited 1×
“The advantages of heterogeneous hypergraphs We argue that the pairwise similarity of heterogeneous hypergraph is a distributed representation [3].”
BackgroundRecognition-by-components: A theory of human image understanding.
Biederman, I. (1987) · Psychological Review · doi:10.1037/0033-295x.94.2.115 · Cited by 5,621 · cited 1×
“Humans can distinguish 30,000 basic object classes [3] and many more subordinate ones (e.g. breeds of dogs).”
BackgroundLearning Deep Architectures for AI
Bengio, Y. (2009) · now publishers, Inc. eBooks · doi:10.1561/9781601982957 · Cited by 4,941
Semi-Supervised Learning Literature Survey
Zhu, X. (2005) · Minds at UW (University of Wisconsin) · Cited by 4,505
DeCAF: A Deep Convolutional Activation Feature for Generic Visual Recognition
Donahue, J., Jia, Y., Vinyals, O., Hoffman, J., Zhang, N., Tzeng, E., & Darrell, T. (2013) · arXiv · doi:10.48550/arxiv.1310.1531 · Cited by 4,019 · cited 1×
“In addition, the Decaf (D) [10] feature is also considered for AwA.”
Background
Research works printed but not linked: 5
- [39] E. Rosch. Classification of real-world objects: Origins and representations in cognition. Thinking: Readings in Cognitive Science , 1977. -
- [32] T. Mikolov, K. Chen, G. Corrado, and J. Dean. Efficient estimation of word representation in vector space. In Proceedings of Workshop at ICLR , 2013. -
- [33] M. Palatucci, G. Hinton, D. Pomerleau, and T. M. Mitchell. Zeroshot learning with semantic output codes. In NIPS , 2009. -
- [34] D. Parikh and K. Grauman. Relative attributes. In ICCV , 2011. -
- [21] Y. Huang, Q. Liu, and D. Metaxas. Video object segmentation by hypergraph cut. In CVPR , 2009. -
Cited by
Items 101–200 of 512
Zero-shot learning and its applications from autonomous vehicles to COVID-19 diagnosis: A review
Rezaei, M., & Shahidi, M. (2020) · Intelligence-Based Medicine · doi:10.1016/j.ibmed.2020.100005 · Cited by 72
“in [20], [21], [22], [23], [24], [25], [13], [16], [17], [18], [26], [27], [19], [28].”
InfluentialGeneralized Zero-Shot Learning With Multiple Graph Adaptive Generative Networks
Xie, G.-S., Zhang, Z., Liu, G., Zhu, F., Liu, L., Shao, L., & Li, X. (2021) · IEEE Transactions on Neural Networks and Learning Systems · doi:10.1109/tnnls.2020.3046924 · Cited by 55
“Furthermore, when leveraging unseen class data for model training, we obtain the transductive ZSL setting [36], [37].”
MethodCompositional Zero-Shot Learning via Fine-Grained Dense Feature Composition
Huynh, D., & Elhamifar, E. (2020) · neural information processing systems · Cited by 97
Efficient Anchor Graph Factorization for Multi-View Clustering
Li, J., Wang, Q., Yang, M., Gao, Q., & Gao, X. (2023) · IEEE Transactions on Multimedia · doi:10.1109/tmm.2023.3340095 · Cited by 52
“(cid:2) AWA [50] is a dataset of animal images, which consisted of 30475 images of 50 animals categories with six feature representationsforeachimage.”
BackgroundSAM: Self-augmentation mechanism for COVID-19 detection using chest X-ray images
Muhammad, U., Hoque, M. Z., Oussalah, M., Keskinarkaus, A., Seppänen, T., & Sarder, P. (2022) · Knowledge-Based Systems · doi:10.1016/j.knosys.2022.108207 · Cited by 54
Paul, A., Shen, T. C., Lee, S., Balachandar, N., Peng, Y., Lu, Z., & Summers, R. M. (2021) · IEEE Transactions on Medical Imaging · doi:10.1109/tmi.2021.3054817 · Cited by 55
Zero-Shot Learning via Robust Latent Representation and Manifold Regularization
Meng, M., & Yu, J. (2018) · IEEE Transactions on Image Processing · doi:10.1109/tip.2018.2881926 · Cited by 48
A Variational Autoencoder with Deep Embedding Model for Generalized Zero-Shot Learning
Ma, P., & Hu, X. G. (2020) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v34i07.6844 · Cited by 51
“Since we don’t use the original image features of the unseen classes during training, our method is inductive rather than trans-ductive (Fu et al. 2015; Guo et al. 2018).”
BackgroundMethodDual VAEGAN: A generative model for generalized zero-shot learning
Luo, Y., Wang, X., & Pourpanah, F. (2021) · Applied Soft Computing · doi:10.1016/j.asoc.2021.107352 · Cited by 52
Guo, J., & Guo, S. (2020) · IEEE Transactions on Multimedia · doi:10.1109/tmm.2020.2984091 · Cited by 56
“…the domain shift problem, including inductive learning -based methods, which enforce additional constraints from the training data [25], [42], and transductive learning - based methods, which assume that the unseen class examples (unlabeled) are also available during training [29], [45], [46].”
Transductive Zero-Shot Learning With Adaptive Structural Embedding
Yu, Y., Ji, Z., Guo, J., & Pang, Y. (2017) · IEEE Transactions on Neural Networks and Learning Systems · doi:10.1109/tnnls.2017.2753852 · Cited by 47
“en and unseen categories are different and potentially unrelated, the embedding is biased when it is directly applied to the unseen data. This is the well-known projection domain shift problem in ZSL [9], [21]. To alleviate this bias, many transductive ZSL approaches have been developed to aggregate the unseen data to…”
BackgroundInfluentialPerson Search by Text Attribute Query As Zero-Shot Learning
Dong, Q., Zhu, X., & Gong, S. (2019) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv.2019.00375 · Cited by 50
“In this work, for the first time we formulate the problem of text attribute query person search as a zero-shot learning (ZSL) problem [35, 10].”
Self-citationSemantic Feature Augmentation in Few-shot Learning.Preprint
Chen, Z., Fu, Y., Zhang, Y., Jiang, Y., Xue, X., & Sigal, L. (2018) · arXiv (Cornell University) · Cited by 52
Self-citationZero Shot Learning via Multi-scale Manifold Regularization
Deutsch, S., Kolouri, S., Kim, K., Owechko, Y., & Soatto, S. (2017) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr.2017.562 · Cited by 51
Joint Learning of Semantic and Latent Attributes
Peng, P., Tian, Y., Xiang, T., Wang, Y., & Huang, T. (2016) · Lecture notes in computer science · doi:10.1007/978-3-319-46493-0_21 · Cited by 48
“…learning mid-level representation from low-level features for object recognition, typically at the fine-grained [6] or instance-level [11]; (3) zero-shot learning where given an attribute ‘prototype’ [12], unseen classes can be recognised by comparing the prototypes with the predicted attributes.”
BackgroundSelf-citationDomain-Invariant Projection Learning for Zero-Shot RecognitionPreprint
Zhao, A., Ding, M., Guan, J., Lu, Z., Xiang, T., & Wen, J.-R. (2018) · arXiv (Cornell University) · doi:10.48550/arxiv.1810.08326 · Cited by 53
“…to whether the predicted labels of the test images are iteratively used for model learning, existing transductive ZSL models fall into two categories: (1) The first category [16, 17, 26, 45, 64] first constructs a graph in the semantic space and then transfers to the test set by label propagation.”
BackgroundSelf-citationGraph and Autoencoder Based Feature Extraction for Zero-shot Learning
Liu, Y., Xie, D., Gao, Q., Han, J., Wang, S., & Gao, X. (2019) · International Joint Conference on Artificial Intelligence · doi:10.24963/ijcai.2019/421 · Cited by 45
Bian, C., Yuan, C., Ma, K., Yu, S., Wei, D., & Zheng, Y. (2021) · IEEE Transactions on Medical Imaging · doi:10.1109/tmi.2021.3131245 · Cited by 48
“First, it is widely accepted that the projection domain shift problem [57], [58] is the fundamental problem in ZSL.”
BackgroundZero-Shot Human Activity Recognition Using Non-Visual Sensors
Machot, F. A., Elkobaisi, M. R., & Kyamakya, K. (2020) · Sensors · doi:10.3390/s20030825 · Cited by 52
“In particular, we refer here to the zero-shot learning, where the classes covered by training and testing samples are disjoint [13,14].”
BackgroundA Discriminative Cross-Aligned Variational Autoencoder for Zero-Shot Learning
Liu, Y., Gao, X., Han, J., & Shao, L. (2022) · IEEE Transactions on Cybernetics · doi:10.1109/tcyb.2022.3164142 · Cited by 45
“The projection domain bias or domain shift problem [10] is another interesting phenomenon in ZSL.”
BackgroundDeep Ranking for Image Zero-Shot Multi-Label Classification
Ji, Z., Cui, B., Li, H., Jiang, Y., Xiang, T., Hospedales, T. M., & Fu, Y. (2020) · IEEE Transactions on Image Processing · doi:10.1109/tip.2020.2991527 · Cited by 42
“Existing transductive approaches generally exploit information of the unseen class instances by exploiting manifolds [7], or self-paced learning [29].”
BackgroundSelf-citationTowards Zero-Shot Learning: A Brief Review and an Attention-Based Embedding Network
Xie, G.-S., Zhang, Z., Xiong, H., Shao, L., & Li, X. (2022) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2022.3208071 · Cited by 43
“However, in addition to seen images, if we further leverage the unseen data (without labels) for training, we get the transductive ZSL/GZSL settings [119], [120].”
BackgroundScalable Zero-Shot Learning via Binary Visual-Semantic Embeddings
Shen, F., Zhou, X. S., Yu, J., Yang, Y., Liu, L., & Shen, H. T. (2019) · IEEE Transactions on Image Processing · doi:10.1109/tip.2019.2899987 · Cited by 44
Attribute-Modulated Generative Meta Learning for Zero-Shot Learning
Li, Y., Liu, Z., Yao, L., & Chang, X. (2021) · IEEE Transactions on Multimedia · doi:10.1109/tmm.2021.3139211 · Cited by 44
“However, this strategy suffers from the domain shift problem, due to distribution differences [21].”
BackgroundBi-Directional Distribution Alignment for Transductive Zero-Shot Learning
Wang, Z., Hao, Y., Mu, T., Li, O., Wang, S., & He, X. (2023) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr52729.2023.01905 · Cited by 52
Cross-Lingual Unsupervised Sentiment Classification with Multi-View Transfer Learning
Fei, H., & Li, P. (2020) · Proceedings of the conference - Association for Computational Linguistics. Meeting · doi:10.18653/v1/2020.acl-main.510 · Cited by 39
“Our work is particularly inspired by Fu et al. (2015) and Zhang et al.”
BackgroundMethodMicroplastic pollution assessment with digital holography and zero-shot learning
Zhu, Y., Lo, H. K. A., Yeung, C. H., & Lam, E. Y. M. (2022) · APL Photonics · doi:10.1063/5.0093439 · Cited by 42
EGANS: Evolutionary Generative Adversarial Network Search for Zero-Shot Learning
Chen, S., Chen, S., Hou, W., Ding, W., & You, X. (2023) · IEEE Transactions on Evolutionary Computation · doi:10.1109/tevc.2023.3307245 · Cited by 40
“, attribute descriptions, sentence embeddings) [1]–[6], i.”
Cross-Modal Zero-Shot-Learning for Tactile Object Recognition
Liu, H., Sun, F., Fang, B., & Guo, D. (2018) · IEEE Transactions on Systems Man and Cybernetics Systems · doi:10.1109/tsmc.2018.2818184 · Cited by 39
“Fu et al. [11] and Zhang and Saligrama [39] searched the new spaces where the feature descriptors and semantic annotations exhibit the maximum similarities for samples of the same class.”
BackgroundPseudo Transfer with Marginalized Corrupted Attribute for Zero-shot Learning
Long, T., Xu, X., Li, Y., Shen, F., Song, J., & Shen, H. T. (2018) · ACM International Conference on Multimedia (ACM MM) · doi:10.1145/3240508.3240715 · Cited by 39
“Besides, to make the most of Xu , transductive ZSL [9, 10] also minimizes LT (Xs ,Xu ) — the discrepancy between φ (Xs ) and φ (Xu ).”
BackgroundSubpixel-Pixel-Superpixel-Based Multiview Active Learning for Hyperspectral Images Classification
Li, Y., Lu, T., & Li, S. (2020) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2020.2971081 · Cited by 38
“…diverse information for model learning, multiview learning has been explored extensively, including multiview supervised methods [23], multiview semisupervised meth-ods [24], multiview AL (MAL)-based methods [25], multiview ensemble methods [26], and multiview transfer learning-based methods [27].”
MethodManifold regularized cross-modal embedding for zero-shot learning
Ji, Z., Yu, Y., Pang, Y., Guo, J., & Zhang, Z. M. (2016) · Information Sciences · doi:10.1016/j.ins.2016.10.025 · Cited by 38
“Further, since the training classes and testing classes are disjoint, directly applying the embedding function learned from training classes to testing ones may cause the domain-shift problem [9].”
BackgroundZero-Shot Learning With Attribute Selection
Guo, Y., Ding, G., Han, J., & Tang, S. (2018) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v32i1.12251 · Cited by 37
“2015b) or both (Norouzi et al. 2013; Fu et al. 2015a), the key building block in ZSL, the attributes themselves, does not seem to receive comparable attention.”
BackgroundZero-shot Image Recognition Using Relational Matching, Adaptation and Calibration
Das, D., & Lee, C. S. G. (2019) · International Joint Conference on Neural Networks (IJCNN) · doi:10.1109/ijcnn.2019.8852315 · Cited by 41
“[18] identified the domain-shift problem and used multiple semantic information sources and label propagation on unlabeled data from the unseen categories to counter the problem.”
BackgroundJoint Visual and Semantic Optimization for zero-shot learning
Wu, H., Yan, Y., Chen, S., Huang, X., Wu, Q., & Ng, M. K. (2021) · Knowledge-Based Systems · doi:10.1016/j.knosys.2021.106773 · Cited by 36
Alleviating Domain Shift via Discriminative Learning for Generalized Zero-Shot Learning
Ye, Y., He, Y., Pan, T., Li, J., & Shen, H. T. (2021) · IEEE Transactions on Multimedia · doi:10.1109/tmm.2021.3063616 · Cited by 39
A survey on visual transfer learning using knowledge graphs
Monka, S., Halilaj, L., & Rettinger, A. (2022) · Semantic Web · doi:10.3233/sw-212959 · Cited by 42
“Knowledge Graph as a Peer Hybrid Transformation Model [117,155,167], [6,17,38,64,133] [155] [68,81,132,158]”
BackgroundMethodMFF: Multi-modal feature fusion for zero-shot learning
Cao, W., Wu, Y., Huang, C., Patwary, M. J. A., & Wang, X. (2022) · Neurocomputing · doi:10.1016/j.neucom.2022.09.070 · Cited by 36
Multimodal zero-shot learning for tactile texture recognition
Cao, G., Jiang, J., Bollegala, D., Li, M., & Luo, S. (2024) · Robotics and Autonomous Systems · doi:10.1016/j.robot.2024.104688 · Cited by 41
A Joint Label Space for Generalized Zero-Shot Classification
Li, J., Lan, X., Long, Y., Liu, Y., Chen, X., Shao, L., & Zheng, N. (2020) · IEEE Transactions on Image Processing · doi:10.1109/tip.2020.2986892 · Cited by 37
“(11) and (12) induce the domain shift problem [43].”
BackgroundDemertzis, K., & Iliadis, L. (2020) · Algorithms · doi:10.3390/a13030061 · Cited by 33
“The ZsL method was adopted based on a heuristic [24], hierarchical parameter search methodology [25].”
MethodVMAN: A Virtual Mainstay Alignment Network for Transductive Zero-Shot Learning
Xie, G.-S., Zhang, X.-Y., Yao, Y., Zhang, Z., Zhao, F., & Shao, L. (2021) · IEEE Transactions on Image Processing · doi:10.1109/tip.2021.3070231 · Cited by 35
“Transductive ZSL: In addition to seen images, utilizing unseen images (without labels) during the training phase yields a transductive setting for ZSL/GZSL [21], termed as TZSL.”
BackgroundMethodZero-Shot Learning Posed as a Missing Data Problem
Zhao, B., Wu, B., Wu, T., & Wang, Y. (2017) · IEEE International Conference on Computer Vision Workshops · doi:10.1109/iccvw.2017.310 · Cited by 35
Learning Aligned Cross-Modal Representation for Generalized Zero-Shot Classification
Fang, Z., Zhu, X., Yang, C. H., Han, Z., Qin, J., & Yin, X.-C. (2022) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v36i6.20614 · Cited by 35
Beyond Semantic Attributes: Discrete Latent Attributes Learning for Zero-Shot Recognition
Qin, J., Wang, Y., Liu, L., Chen, J., & Shao, L. (2016) · IEEE Signal Processing Letters · doi:10.1109/lsp.2016.2612247 · Cited by 33
Transductive Learning With Prior Knowledge for Generalized Zero-Shot Action Recognition
Su, T., Wang, H., Qi, Q., Wang, L., & He, B. (2023) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2023.3284977 · Cited by 35
“On the other hand, the ZSL methods can be divided into two categories: (1) inductive ZSL which only adopts labeled source data for model learning, and (2) transductive ZSL where unlabeled target data are also adopted for model training [8], [9].”
BackgroundMethodIndicative Vision Transformer for end-to-end zero-shot sketch-based image retrieval
Zhang, H., Cheng, D., Kou, Q., Asad, M., & Jiang, H. (2024) · Advanced Engineering Informatics · doi:10.1016/j.aei.2024.102398 · Cited by 31
Characterizing Word Embeddings for Zero-Shot Sensor-Based Human Activity Recognition
Matsuki, M., Lago, P., & Inoue, S. (2019) · Sensors · doi:10.3390/s19225043 · Cited by 38
Image-Specific Classification With Local and Global Discriminations
Zhang, C., Cheng, J., Li, C., & Tian, Q. (2017) · IEEE Transactions on Neural Networks and Learning Systems · doi:10.1109/tnnls.2017.2748952 · Cited by 31
Visual-Semantic Aligned Bidirectional Network for Zero-Shot Learning
Gao, R., Hou, X., Qin, J., Shen, Y., Long, Y., Liu, L., Zhang, Z., & Shao, L. (2022) · IEEE Transactions on Multimedia · doi:10.1109/tmm.2022.3145666 · Cited by 32
“Second, the inter-domain shift [9] is also a critical issue that hinders the over-all ZSL performance, which has not been explicitly mitigated in existing bidirectional methods [24], [26].”
BackgroundLearning cross-domain semantic-visual relationships for transductive zero-shot learning
Lv, F., Zhang, J., Yang, G., Feng, L., Yu, Y., & Duan, L. (2023) · Pattern Recognition · doi:10.1016/j.patcog.2023.109591 · Cited by 30
“Transductive ZSL aims to relieve the projection shift between the source and target categories through introducing unlabeled images from novel classes during training [7].”
BackgroundInfluentialHierarchical Prototype Learning for Zero-Shot Recognition
Zhang, X., Gui, S., Zhu, Z., Zhao, Y., & Liu, J. (2019) · IEEE Transactions on Multimedia · doi:10.1109/tmm.2019.2959433 · Cited by 31
“It has been proved in [21] that transductive ZSL models can indeed improve generalisation accuracy compared with inductive models.”
InfluentialGraph-Based Visual-Semantic Entanglement Network for Zero-Shot Image Recognition
Hu, Y., Wen, G., Chapman, A. P., Yang, P., Luo, M., Xu, Y., Dai, D., & Hall, W. (2021) · IEEE Transactions on Multimedia · doi:10.1109/tmm.2021.3082292 · Cited by 33
What Can Knowledge Bring to Machine Learning?—A Survey of Low-shot Learning for Structured Data
Hu, Y., Chapman, A. P., Wen, G., & Hall, D. W. (2022) · ACM Transactions on Intelligent Systems and Technology · doi:10.1145/3510030 · Cited by 36
Boosting Generative Zero-Shot Learning by Synthesizing Diverse Features with Attribute Augmentation
Zhao, X., Shen, Y., Wang, S., & Zhang, H. (2022) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v36i3.20256 · Cited by 36
A Deep Learning Framework for Tactile Recognition of Known as Well as Novel Objects
Abderrahmane, Z., Ganesh, G., Crosnier, A., & Cherubini, A. (2019) · IEEE Transactions on Industrial Informatics · doi:10.1109/tii.2019.2898264 · Cited by 30
“First, the recognition of novel objects is limited by the domain shift problem [49] and the correlation between”
BackgroundEnergy Confused Adversarial Metric Learning for Zero-Shot Image Retrieval and Clustering
Chen, B., & Deng, W. (2019) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v33i01.33018134 · Cited by 29
Transductive Zero-Shot Action Recognition via Visually Connected Graph Convolutional Networks
Xu, Y., Han, C., Qin, J., Xu, X., Han, G., & He, S. (2020) · IEEE Transactions on Neural Networks and Learning Systems · doi:10.1109/tnnls.2020.3015848 · Cited by 33
“There are many approaches [2], [4], [7], [52], [13] in ZSL rely on various attributes to represent categories.”
BackgroundA Novel Zero-Shot Learning Method With Feature Generation for Intelligent Fault Diagnosis
Liao, W., Wu, L., Xu, S., & Fujimura, S. (2025) · IEEE Transactions on Industrial Informatics · doi:10.1109/tii.2025.3526478 · Cited by 33
Multi-domain few-shot image recognition with knowledge transfer
Li, M., Wang, R., Yang, J., Xue, L., & Hu, M. Q. (2021) · Neurocomputing · doi:10.1016/j.neucom.2021.01.123 · Cited by 28
Hardness Sampling for Self-Training Based Transductive Zero-Shot Learning
Liu, B., Dong, Q., & Hu, Z. (2021) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr46437.2021.01623 · Cited by 37
“[6] proposed a method where visual features and semantic features were projected into a multi-view embedding space and then a hyper-graph was constructed using the unlabeled data for label propagation.”
MethodResidual Tuning: Toward Novel Category Discovery Without Labels
Liu, Y., & Tuytelaars, T. (2022) · IEEE Transactions on Neural Networks and Learning Systems · doi:10.1109/tnnls.2022.3140235 · Cited by 39
“Alternatively, the work in [12] addressed ZSD under a transductive setting [13] where the unseen images can be used as well during training.”
Zero-Shot Learning With Attentive Region Embedding and Enhanced Semantics
Liu, Y., Dang, Y., Gao, X., Han, J., & Shao, L. (2022) · IEEE Transactions on Neural Networks and Learning Systems · doi:10.1109/tnnls.2022.3202014 · Cited by 32
“Thus, the supervised classifier can alleviate the projection domain bias problem [12] with the synthetic unseen samples.”
BackgroundAMP 0 : Species-Specific Prediction of Anti-microbial Peptides Using Zero and Few Shot Learning
Gull, S., & Minhas, F. (2020) · IEEE Transactions on Computational Biology and Bioinformatics · doi:10.1109/tcbb.2020.2999399 · Cited by 28
“tions for novel classes which are not available at training time [31], [32], [33].”
BackgroundMethodPrototype Completion for Few-Shot Learning
Zhang, B., Li, X., Ye, Y., & Feng, S. (2023) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2023.3277881 · Cited by 30
Towards zero-shot learning generalization via a cosine distance loss
Pan, C., Huang, J. X., Hao, J., & Gong, J. (2019) · Neurocomputing · doi:10.1016/j.neucom.2019.11.011 · Cited by 27
Semantically Consistent Regularization for Zero-Shot RecognitionPreprint
Morgado, P. M., & Vasconcelos, N. M. (2017) · arXiv (Cornell University) · doi:10.48550/arxiv.1704.03039 · Cited by 26
“Zero-shot learning Most current solutions to ZSL fall under two main categories: RIS and RULE.”
BackgroundCohesion Intensive Hash Code Book Coconstruction for Efficiently Localizing Sketch Depicted Scenes
Fang, Y., Li, P., Zhang, J., & Ren, P. (2021) · IEEE Transactions on Geoscience and Remote Sensing · doi:10.1109/tgrs.2021.3132296 · Cited by 25
“Fu et al. [14] proposed a framework to rectify the biased projections in a multi-view embedding space to solve the problem of projection domain shift.”
BackgroundGraph Knows Unknowns: Reformulate Zero-Shot Learning as Sample-Level Graph Recognition
Guo, J., Guo, S., Zhou, Q., Liu, Z., Lu, X., & Huo, F. (2023) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v37i6.25942 · Cited by 25
“This issue brings about the domain bias problem (Fu et al. 2015) when generalizing the learned knowledge from seen to unseen classes, i.”
BackgroundJoint Dictionaries for Zero-Shot Learning
Kolouri, S., Rostami, M. G., Owechko, Y., & Kim, K. (2018) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v32i1.11649 · Cited by 30
Query Attack via Opposite-Direction Feature:Towards Robust Image RetrievalPreprint
Zheng, Z., Zheng, L., Yang, Y., & Wu, F. (2018) · arXiv (Cornell University) · doi:10.48550/arxiv.1809.02681 · Cited by 27
“In some early works [14, 6, 5, 2], the intermediate semantic representation is usually learned from the source dataset and applied to the target dataset.”
BackgroundBias-Eliminated Semantic Refinement for Any-Shot Learning
Feng, L., Zhao, C., & Li, X. (2022) · IEEE Transactions on Image Processing · doi:10.1109/tip.2022.3152631 · Cited by 24
“[47] combined multiple semantic embeddings and performed hypergraph label propagation.”
Yang, B., & Sun, H. (2022) · Journal of the Brazilian Society of Mechanical Sciences and Engineering · doi:10.1007/s40430-022-03965-2 · Cited by 24
“But they predominantly suffer from the projection domain shift problem [20].”
BackgroundAlignment-Uniformity aware Representation Learning for Zero-shot Video Classification
Pu, S., Zhao, K., & Zheng, M. (2022) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr52688.2022.01934 · Cited by 36
“However, these methods learn associated representations within limited classes, thus facing the following two critical problems [11, 13]: (1) semantic-gap: manifolds inconsistency between visual and semantics features, and (2)”
BackgroundInfluentialAdaptive Metric Learning For Zero-Shot Recognition
Jiang, H., Wang, R., Shan, S., & Chen, X. (2019) · IEEE Signal Processing Letters · doi:10.1109/lsp.2019.2917148 · Cited by 25
Zero-shot learning by mutual information estimation and maximization
Tang, C., Yang, X., Lv, J., & He, Z. (2020) · Knowledge-Based Systems · doi:10.1016/j.knosys.2020.105490 · Cited by 27
Cross Knowledge-based Generative Zero-Shot Learning approach with Taxonomy Regularization
Xie, C., Xiang, H., Zeng, T., Yang, Y., Yu, B., & Liu, Q. (2021) · Neural Networks · doi:10.1016/j.neunet.2021.02.009 · Cited by 25
Unifying Visual Attribute Learning with Object Recognition in a Multiplicative Framework
Liang, K., Chang, H., Ma, B., Shan, S., & Chen, X. (2018) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2018.2836461 · Cited by 23
Vocabulary-Informed Zero-Shot and Open-Set Learning
Fu, Y., Wang, X., Dong, H., Jiang, Y., Wang, M., Xue, X., & Sigal, L. (2019) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2019.2922175 · Cited by 26
“tic space using Support Vector Regressors (SVR) [18], [48] or neural network [75]; (2) projecting visual features and semantic entities into a common new space, such as SJE [3], WSABIE [85], ALE [1], DeViSE [24], and CCA [26], [29]; (3) learning the embeddings by regressing from the semantic space to visual features,…”
MethodSelf-citationMetAdapt: Meta-learned task-adaptive architecture for few-shot classification
Doveh, S., Schwartz, E., Xue, C., Feris, R., Bronstein, A. M., Giryes, R., & Karlinsky, L. (2021) · Pattern Recognition Letters · doi:10.1016/j.patrec.2021.05.010 · Cited by 23
Transferrable Feature and Projection Learning with Class Hierarchy for Zero-Shot Learning
Li, A., Lu, Z., Guan, J., Xiang, T., Wang, L., & Wen, J.-R. (2020) · International Journal of Computer Vision · doi:10.1007/s11263-020-01342-x · Cited by 26
“To overcome the projection domain shift (Fu et al., 2015a) caused by the domain gap in ZSL, transductive ZSL has been proposed to learn the projection function with not only labelled data from seen classes but also unlabelled data from unseen classes.”
BackgroundMethodInfluentialSelf-citationDescribing Unseen Classes by Exemplars: Zero-Shot Learning Using Grouped Simile Ensemble
Long, Y., & Shao, L. (2017) · IEEE Winter Conference on Applications of Computer Vision · doi:10.1109/wacv.2017.106 · Cited by 23
“Many recent approaches adopt such an embedding manner and achieve promising results [13, 4, 33, 15, 7, 19, 39, 8, 23].”
BackgroundA cross-modal deep metric learning model for disease diagnosis based on chest x-ray images
Jin, Y., Lu, H., Zhao, L., & Wang, Y. (2023) · Multimedia Tools and Applications · doi:10.1007/s11042-023-14790-7 · Cited by 23
Zero-shot classification by transferring knowledge and preserving data structure
Li, X., Fang, M., & Wu, J. (2017) · Neurocomputing · doi:10.1016/j.neucom.2017.01.038 · Cited by 24
Multi-View Nonparametric Discriminant Analysis for Image Retrieval and Recognition
Cao, G., Iosifidis, A., & Gabbouj, M. (2017) · IEEE Signal Processing Letters · doi:10.1109/lsp.2017.2748392 · Cited by 22
“We also demonstrate the effectiveness of multi-view embeddings in tackling the domain shift problem for zero-shot recognition [24].”
BackgroundZero-shot image classification using coupled dictionary embedding
Rostami, M. G., Kolouri, S., Murez, Z., Owechko, Y., Eaton, E., & Kim, K. (2022) · Machine Learning with Applications · doi:10.1016/j.mlwa.2022.100278 · Cited by 31
Deep Learning for Free-Hand Sketch: A Survey and A ToolboxPreprint
Xu, P., Hospedales, T. M., Yin, Q., Song, Y.-Z., Xiang, T., & Wang, L. (2020) · arXiv (Cornell University) · Cited by 30
Self-citationAdjust to reality: LLM-driven test-time semantic adjustment for zero-shot fault diagnosis
Zhao, J., Yue, J., Zhao, C., & Chen, C. (2025) · Control Engineering Practice · doi:10.1016/j.conengprac.2025.106406 · Cited by 23
From Pixel to Patch: Synthesize Context-Aware Features for Zero-Shot Semantic Segmentation
Gu, Z., Zhou, S., Niu, L., Zhao, Z., & Zhang, L. (2022) · IEEE Transactions on Neural Networks and Learning Systems · doi:10.1109/tnnls.2022.3145962 · Cited by 24
“Many existing ZSL methods [2]–[4], [12], [46]–[49] attempted to learn a mapping between feature space and semantic space.”
Generative Meta-Adversarial Network for Unseen Object Navigation
Zhang, S., Li, W., Song, X., Bai, Y., & Jiang, S. (2022) · Lecture notes in computer science · doi:10.1007/978-3-031-19842-7_18 · Cited by 33
“Feature generation for unseen objects has been widely studied in zero-shot recognition tasks [23,1,17,16,47,46,58].”
BackgroundMulti-view enhanced zero-shot node classification
Wang, J., Wu, L., Zhao, H., & Jia, N. (2023) · Information Processing & Management · doi:10.1016/j.ipm.2023.103479 · Cited by 23
Prototype-Augmented Self-Supervised Generative Network for Generalized Zero-Shot Learning
Wu, J., Zhang, T., Zha, Z.-J., Luo, J., Zhang, Y., & Wu, F. H. (2024) · IEEE Transactions on Image Processing · doi:10.1109/tip.2024.3351439 · Cited by 23
“In this paper, we apply GZSL in a transductive way, assuming that unlabeled target data is accessible during the training stage [23].”
BackgroundMethodComputational Social Indicators
Feng, F., Nie, L., Wang, X., Hong, R., & Chua, T. (2017) · International ACM SIGIR Conference on Research and Development in Information Retrieval (SIGIR) · doi:10.1145/3077136.3080773 · Cited by 21
“Typically, multiple views are mapped into a common space by different algorithms [40], including canonical correlation analysis [15], dictionary learning [4], matrix factorization [20, 29, 41].”
BackgroundMethodObject Categorization Using Class-Specific Representations
Zhang, C., Cheng, J., Li, L., Li, C., & Tian, Q. (2017) · IEEE Transactions on Neural Networks and Learning Systems · doi:10.1109/tnnls.2017.2757497 · Cited by 21
Sketch-a-Classifier: Sketch-Based Photo Classifier Generation
Hu, C., Li, D., Song, Y.-Z., Xiang, T., & Hospedales, T. M. (2018) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr.2018.00952 · Cited by 26
Self-citationOpen Set Adversarial ExamplesPreprint
Zheng, Z., Zheng, L., Hu, Z., & Yang, Y. (2018) · arXiv (Cornell University) · Cited by 22
Deep Semantic Structural Constraints for Zero-Shot Learning
Li, Y., Jia, Z., Zhang, J., Huang, K., & Tan, T. N. (2018) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v32i1.12244 · Cited by 20
“…Zero-Shot Leanring (ZSL) into image classification (Lampert, Nickisch, and Harmeling 2009; Frome et al. 2013; Norouzi et al. 2013; Socher et al. 2013; Fu et al. 2015; Akata et al. 2015; Romera-Paredes and Torr 2015; Bucher, Herbin, and Jurie 2016; Akata et al. 2016; Huang, Loy, and Tang 2016;…”
BackgroundAdversarial Zero-shot Learning With Semantic Augmentation
Tong, B., Klinkigt, M., Chen, J., Cui, X., Kong, Q., Murakami, T., & Kobayashi, Y. (2018) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v32i1.11886 · Cited by 22
“2015b), SS-VOC (Fu et al. 2015a), PDDM (Huang, Loy, and Tang 2016).”
BackgroundInference guided feature generation for generalized zero-shot learning
Han, Z., Fu, Z., Li, G., & Yang, J. J. (2020) · Neurocomputing · doi:10.1016/j.neucom.2020.10.080 · Cited by 20
Zero-Shot Chinese Text Recognition via Matching Class Embedding
Huang, Y., Jin, L., & Peng, D. (2021) · Lecture notes in computer science · doi:10.1007/978-3-030-86334-0_9 · Cited by 24
Metadata
- Authors
- Fu, Y., Hospedales, T. M., Xiang, T., & Gong, S.
- Affiliations
- Walt Disney (United States) (US)
- Queen Mary University of London (GB)
- Venue
- IEEE Transactions on Pattern Analysis and Machine Intelligence
- Published
- 2015-11-01
- Volume
- 37
- Issue
- 11
- Pages
- 2332-2345
- ISSN
- 0162-8828 (ISSN-L)
- 2160-9292
- OpenAlex
- W2141350700
- Cited by
- 514
- References
- 74
- Citations by year
- 2026: 17 · 2025: 32 · 2024: 32 · 2023: 52 · 2022: 60 · 2021: 75 · 2020: 62 · 2019: 57 · 2018: 64 · 2017: 41 · 2016: 21 · 2015: 1
- FWCI
- 41.51
- Type
- journal article
- Language
- en
- Open access
- green
- arxiv.org
- Locations
- IEEE Transactions on Pattern Analysis and Machine Intelligence (publishedVersion)
- PubMed (publishedVersion)
- arXiv (Cornell University) (submittedVersion)
- Queen Mary Research Online (Queen Mary University of London) (submittedVersion)
- arXiv (Cornell University)
- Topics
- Physical Sciences › Computer Science › Artificial Intelligence › Domain Adaptation and Few-Shot Learning
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Advanced Neural Network Applications
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Advanced Image and Video Retrieval Techniques
- Keywords
- zero-shot learning, transductive learning, multi-view embedding, semantic representation, transfer learning, manifold structure
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.